Discovery of potent inhibitors of α-synuclein aggregation using structure-based iterative learning

Area of Science:

• Computational chemistry
• Neuroscience
• Drug discovery

Background:

• Drug discovery for neurodegenerative diseases, particularly Parkinson's disease, faces high costs and failure rates.
• Alpha-synuclein aggregation is a critical pathological process in Parkinson's disease and other synucleinopathies.
• Conventional drug discovery pipelines are inefficient for developing disease-modifying therapies.

Purpose of the Study:

• To develop a machine learning approach for identifying small molecule inhibitors of alpha-synuclein aggregation.
• To target the autocatalytic secondary nucleation process involved in alpha-synuclein aggregate proliferation.
• To discover compounds that bind to the catalytic sites on alpha-synuclein aggregates.

Main Methods:

• Utilized structure-based machine learning in an iterative framework.
• Employed a strategy to identify and progressively optimize secondary nucleation inhibitors.
• Focused on small molecules targeting the catalytic sites of alpha-synuclein aggregates.

Main Results:

• Successfully identified small molecule inhibitors of alpha-synuclein aggregation.
• The machine learning approach facilitated the discovery of potent compounds.
• Achieved a two-orders-of-magnitude increase in compound potency compared to previous reports.

Conclusions:

• Machine learning offers a promising strategy to reduce costs and failure rates in drug discovery.
• The developed approach effectively identifies potent inhibitors of alpha-synuclein secondary nucleation.
• This method holds potential for accelerating the development of therapeutics for Parkinson's disease and related disorders.